Next Article in Journal
Pollen Grain Classification Using Some Convolutional Neural Network Architectures
Previous Article in Journal
GOYA: Leveraging Generative Art for Content-Style Disentanglement
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Toward Unbiased High-Quality Portraits through Latent-Space Evaluation

by
Doaa Almhaithawi
1,*,
Alessandro Bellini
2 and
Tania Cerquitelli
1
1
Department of Control and Computer Engineering, Politecnico di Torino, 10129 Torino, Italy
2
Prime Lab, Mathema s.r.l., 50142 Florence, Italy
*
Author to whom correspondence should be addressed.
J. Imaging 2024, 10(7), 157; https://doi.org/10.3390/jimaging10070157
Submission received: 14 May 2024 / Revised: 21 June 2024 / Accepted: 26 June 2024 / Published: 28 June 2024
(This article belongs to the Topic Applications in Image Analysis and Pattern Recognition)

Abstract

Images, texts, voices, and signals can be synthesized by latent spaces in a multidimensional vector, which can be explored without the hurdles of noise or other interfering factors. In this paper, we present a practical use case that demonstrates the power of latent space in exploring complex realities such as image space. We focus on DaVinciFace, an AI-based system that explores the StyleGAN2 space to create a high-quality portrait for anyone in the style of the Renaissance genius Leonardo da Vinci. The user enters one of their portraits and receives the corresponding Da Vinci-style portrait as an output. Since most of Da Vinci’s artworks depict young and beautiful women (e.g., “La Belle Ferroniere”, “Beatrice de’ Benci”), we investigate the ability of DaVinciFace to account for other social categorizations, including gender, race, and age. The experimental results evaluate the effectiveness of our methodology on 1158 portraits acting on the vector representations of the latent space to produce high-quality portraits that retain the facial features of the subject’s social categories, and conclude that sparser vectors have a greater effect on these features. To objectively evaluate and quantify our results, we solicited human feedback via a crowd-sourcing campaign. Analysis of the human feedback showed a high tolerance for the loss of important identity features in the resulting portraits when the Da Vinci style is more pronounced, with some exceptions, including Africanized individuals.
Keywords: latent space density; industrial survey; Isometric Mapping (ISOMAP); StyleGAN2; model bias; qualitative and quantitative analysis; dimensionality reduction latent space density; industrial survey; Isometric Mapping (ISOMAP); StyleGAN2; model bias; qualitative and quantitative analysis; dimensionality reduction

Share and Cite

MDPI and ACS Style

Almhaithawi, D.; Bellini, A.; Cerquitelli, T. Toward Unbiased High-Quality Portraits through Latent-Space Evaluation. J. Imaging 2024, 10, 157. https://doi.org/10.3390/jimaging10070157

AMA Style

Almhaithawi D, Bellini A, Cerquitelli T. Toward Unbiased High-Quality Portraits through Latent-Space Evaluation. Journal of Imaging. 2024; 10(7):157. https://doi.org/10.3390/jimaging10070157

Chicago/Turabian Style

Almhaithawi, Doaa, Alessandro Bellini, and Tania Cerquitelli. 2024. "Toward Unbiased High-Quality Portraits through Latent-Space Evaluation" Journal of Imaging 10, no. 7: 157. https://doi.org/10.3390/jimaging10070157

APA Style

Almhaithawi, D., Bellini, A., & Cerquitelli, T. (2024). Toward Unbiased High-Quality Portraits through Latent-Space Evaluation. Journal of Imaging, 10(7), 157. https://doi.org/10.3390/jimaging10070157

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop